From mboxrd@z Thu Jan 1 00:00:00 1970 From: Stephen Hemminger Subject: Re: [PATCH] igb_uio: allow multi-process access Date: Mon, 11 Dec 2017 17:38:55 -0800 Message-ID: <20171211173855.3eed4623@xeon-e3> References: <1512784653-128951-1-git-send-email-xiao.w.wang@intel.com> Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Cc: ferruh.yigit@intel.com, dev@dpdk.org To: Xiao Wang Return-path: Received: from mail-pg0-f67.google.com (mail-pg0-f67.google.com [74.125.83.67]) by dpdk.org (Postfix) with ESMTP id 7AF13201 for ; Tue, 12 Dec 2017 02:38:59 +0100 (CET) Received: by mail-pg0-f67.google.com with SMTP id e14so12287505pgr.9 for ; Mon, 11 Dec 2017 17:38:59 -0800 (PST) In-Reply-To: <1512784653-128951-1-git-send-email-xiao.w.wang@intel.com> List-Id: DPDK patches and discussions List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: dev-bounces@dpdk.org Sender: "dev" On Fri, 8 Dec 2017 17:57:33 -0800 Xiao Wang wrote: > In some case, one device are accessed by different processes via > different BARs, so one uio device may be opened by more than one > process, for this case we just need to enable interrupt once, and > pci_clear_master only when the last process closed. > > Fixes: 5f6ff30dc507 ("igb_uio: fix interrupt enablement after FLR in VM") Yes, this makes sense. > > Signed-off-by: Xiao Wang > --- > lib/librte_eal/linuxapp/igb_uio/igb_uio.c | 7 +++++++ > 1 file changed, 7 insertions(+) > > diff --git a/lib/librte_eal/linuxapp/igb_uio/igb_uio.c b/lib/librte_eal/linuxapp/igb_uio/igb_uio.c > index a3a98c1..c239d98 100644 > --- a/lib/librte_eal/linuxapp/igb_uio/igb_uio.c > +++ b/lib/librte_eal/linuxapp/igb_uio/igb_uio.c > @@ -45,6 +45,7 @@ struct rte_uio_pci_dev { > struct uio_info info; > struct pci_dev *pdev; > enum rte_intr_mode mode; > + uint32_t ref_cnt; Simple unsigned reference count is not SMP safe on all architectures. In kernel it is recommended to use refcount_t and associated API's. Note: refcount_t was introduced in last 2 years and some DPDK users still have ancient kernels. > }; > > static char *intr_mode; > @@ -336,6 +337,9 @@ struct rte_uio_pci_dev { > struct pci_dev *dev = udev->pdev; > int err; > > + if (++(udev->ref_cnt) > 1) > + return 0; Do not use (unnecessary) parenthesis. C precedence order is well defined. > /* set bus master, which was cleared by the reset function */ > pci_set_master(dev); > > @@ -354,6 +358,9 @@ struct rte_uio_pci_dev { > struct rte_uio_pci_dev *udev = info->priv; > struct pci_dev *dev = udev->pdev; > > + if (--(udev->ref_cnt) > 0) > + return 0; > + > /* disable interrupts */ > igbuio_pci_disable_interrupts(udev); >